Artificial Intelligence, Ethics and International Human Rights Law

1. Introduction

Artificial Intelligence (AI) is changing the world and its potential impact on societies is unlimited. It facilitates the use of virtual space for dialogue, cooperation, negotiation, debate, and trade. This space is used for good things but also bad ones such as terrorism and cyber criminality. AI creates new challenges for all actors and fundamentally influences human rights. This invites states, international organizations, researchers, and policymakers to think of new ways, new mechanisms, and new policy actions to protect human rights in the age of AI. The key questions guiding AI Ethics are:

• What are the risks imposed by AI to human rights?
• What are the limits of international law?
• Which, if any, ethical framework is sufficient to protect human rights?
• What are the limits and solutions of AI?

In this paper, these questions are engaged, particularly with regards to the implications of AI on human rights, and the importance of the international instruments of ethics in artificial intelligence.

2. Artificial Intelligence, new risks for human rights

AI has proven the latest 'weapon' to be employed by states and other actors, increasing both current and new risks to national and global security. AI offers the possibility to analyze human behaviors, moods, and beliefs based on available data (Brundage et al). Malicious use of AI exacerbates risk, increasing its negative impact, including its influence on maintaining or disrupting democratic balances. The right to life, freedom, privacy, and universal values and principals are threatened by AI.

2.1. Right to life threatened by autonomous weapons systems

The right to life is central in the debates surrounding the potential impact of AI and autonomous weapon systems. As stated in the Preamble to the Charter of United Nations, "We the people of the United Nations are determined to save succeeding generations from the scourge of war" (United Nation) and the Universal Declaration of Human Rights (UDHR) recognizes in Article 3 that "Everyone has the right to life, liberty, and security of person." But the growing investment in lethal autonomous weapons is a serious threat to this right. Worldwide spending on robotics is expected to reach $87 Billion by 2025 (Meldonet al).

International law aims to ensure peace and security, but AI has imposed new challenges that limit its efficiency. These challenges have been increased by the COVID-19 pandemic, which is facilitating the transition to a new world order different from the global order established after the Second World War.

In the same context, the introduction of autonomous weapons systems (AWS) has created a controversial discussion between states according to Burri because they are real risks to the right to life, which requires an urgent review of the use of force, as cited in the UN Charter. According to Burri "fully autonomous weapons systems (i.e., systems that select and engage targets without meaningful human control) are likely to be banned through a new international legal instrument, while the use of weapons systems equipped with a low level of autonomy will be lawful" (99). Goldstein argued that state competition toward AWS leads us to the assessment that the current trade crisis between China and the US may be escalated to an open military conflict with the use of AI weapons (52). Fully AWS, or as they have been called by the (Human Rights Watch), "human-out-of-the-loop weapons", are currently the most dangerous threat to the right to life, peace, and security. The threat of AWS is "the problem of reaction time, which threatens to turn humans in and on the loop into liabilities" (Leys 51). Anderson and Waxman explain that "such systems are much easier to create than lawful ones (Anderson and Waxman). The speed necessary to respond to such adversary systems in the field, though, might well create demand for defensive systems that feature greater autonomy in decision-making" (Anderson & Waxman 8). First, the future of humanity will not be decided by state actors when AWS is employed. Second, all these new technologies are growing faster than international law as stated by (Anderson and Waxman 8). Thus, international norms such as those concerning the use of force and defense need to be revised to ensure peace and security and protect the right to life.

The legitimacy of new weapons is a culminating point in the controversial discussion about AWS. Article 36 of the 1977 revision of the Geneva Conventions provides that "in the study, development, acquisition or adoption of a new weapon, means or method of warfare, a High Contracting Party is under an obligation to determine whether its employment would, in some or all circumstances, be prohibited by this protocol or by any other rule of international law applicable to the High Contracting Party" (United Nations 258). Davison confirms that the commander (human or otherwise) in the use of weapon systems should respect their core legal obligation (3). Anderson and Waxman also query whether "robots [can] ensure the distinction between military and civilian objects or between active combatants and innocent civilians?" (8). This means that states need to invest more in ethics in artificial intelligence to prevent violations of international humanitarian law. Ethics is the only way to minimize the risks imposed using AI in the military. Thus, states need to collaborate with all stakeholders to ensure the technical and legal protection of human rights in peace and during the war. Rethinking international law and national legislation is now an obligation rather than a choice. In the same context, states need to adapt their policies and international mechanisms alongside evolutions in AWS. They should also review their responsibilities in consideration of the right to life as a significant right threatened by AWS, including during armed conflicts.

2.2. Freedom and Artificial Intelligence

According to Article 19 of the Universal Declaration of Human Rights "Everyone has the right to freedom of opinion and expression, this right includes freedom to hold opinions without interference and to seek, receive and impart information and ideas through any media and regardless of frontiers" (UDHR). The key question is how can we ensure freedom of expression when public opinion is influenced by AI? This new tool is imposing new risks to freedom.

AI offers new tools to create content (audio and visual analyses) and while the possibility exists for AI to support the freedom of expression, which is a cornerstone for democracy and the enemy for corruption, in reality it increases the control of social media platforms and freedom of expression by the government. AI systems contained in social media are also used to influence public opinion and to guide social movements considering workflow optimization, automated content creation, content creation from legacy archives, content selection for targeting audience demographics, optimization of asset selection, metadata creation, and content personalization (ITU). AI can personalize, generate, and filter content. It has terrifying implications for freedom of expression, social movement, and election campaigns. Questions arise concerning non-trusted or fake information published by the media, but which is selected and kept trending by AI. How can we determine the level of trust in media which can be manipulated by governments, advertisers, algorithms or other third parties seeking to persuade users and recipients of such information?

Some AI systems are more efficient than humans in certain tasks such as mimicking others' voices and images to influence people and to create political changes (also known as deep fakes). There is also the concept of machine learning software that creates fake videos (Cole, 2018). This new technology developed by Chinese tech giant Baidu can reproduce a believable fake voice with just 3.7 seconds of audio much as the concept of machine learning software that creates fake videos (Cole, 2018). In the same context, Montreal-based AI start-up Lyrebird claims it can do text-to-speech using just one minute of audio (Cole, 2018). This means that individuals are no longer in control of the creation and security of their public opinion, rather they are represented by AI. They can no longer trust their own autonomy when their expression of thought is influenced by the information spreading on social networks and their personal interactions with algorithms are manipulated and abused.

The use of smartphones becomes a real risk to freedom of expression considering the number of arrests in some countries after individuals posted on Facebook calls for freedom during quarantine, such as the Australian woman Zoe Buhler, who was arrested at her home after she created a "Freedom Day" event on Facebook calling for people to protest against the Victorian government's coronavirus lockdown measures (The Guardian).

In Tunisia, Amna Al-Sharqi was arrested for posting a text on her Facebook page entitled "Surah Corona " (The New Arab). According to Freedom House, "Rather than protecting users, the application of national sovereignty to cyberspace has given authorities free rein to crack down on human rights" (Shabaz & Funk 2). A large number of countries are using COVID-19 to justify their engagement in mass surveillance in partnership with companies and telecommunication providers, the most draconian approach being adopted by China (Shabaz & Funk 2).

The pandemic, and the laws adopted by some governments to face disinformation, has created a serious restriction on freedom and privacy and it creates new legal challenges related to international human rights. For that reason, UNESCO has issued guidelines for judges and courts, both at national and regional levels that can serve as references to apply the theoretical frameworks of international law and human rights standards to protect and promote freedom of expression" (UNESCO a).

2.3. Protecting privacy in the age of Artificial Intelligence

According to Article 12 of the Universal Declaration of Human Rights, "No one shall be subjected to arbitrary interference with his privacy, family, home or correspondence, nor to attacks upon his honor and reputation." However, systems that combine data from satellite imagery, facial recognition-powered cameras, and cell phone location information, etc., can provide a detailed picture of an individual's movements as well as predict future movements and location. It could therefore easily be used by governments to facilitate more precise restriction of the freedom of movement at both the individual and group level and by foreign actors who are targeting political changes (Access Now 21). Voting behaviour and election campaigns are also influenced by social media (Brundage et al 29). We are constantly connected to our smartphones which facilitates the search for each case of Covid-19 to mitigate the impact and magnitude of this pandemic. Today's smartphones even allow remote access to a person's electro-gram. This creates new risks and challenges ranging from privacy to freedom of expression, given the tension between individuals and governments regarding human rights and democracy. Facial recognition is one of the programs which raises privacy issues which can increase digital dictatorship. Conversely, "the facial recognition market is expected to grow to US$7.7 billion in 2022 from US$4 billion in 2017. That is because facial recognition has all kinds of commercial applications. It can be used for everything from surveillance to marketing" (Symanovich). COVID-19 and AI are taking societies around the world to another phase in history with the increased use of robots for online shopping and deliveries, digital and contactless payments, remote work, distance learning, etc. AI is changing our lives and is influencing all sectors, as argued by the (OECD a).

In this context, digital technology can play a role in contact tracing programs implemented in the Member States. Several countries are using artificial intelligence to ensure access to information and to trace COVID-19 but these apps are also tracking individuals according to Freedom House (Shabaz & Funk 2). Member States are obliged under the International Health Regulations, to develop public health surveillance systems that capture critical data for their COVID-19 response, while ensuring that such systems are transparent, responsive to the concerns of communities, and do not impose unnecessary burdens, for example, infringements on privacy" (WHO 1). This is creating serious tension between states. AI creates new challenges for international law regarding human rights, and it can be a risk to freedom and privacy. According to the World Health Organization, "such uses of data may also threaten fundamental human rights and liberties during and after the COVID-19 pandemic. Surveillance can quickly traverse the blurred line between disease surveillance and population surveillance" (WHO 1). Free and open scientific data impose other challenges that necessitate rethinking international law in consideration of the appearance of new notions linked to the states and their sovereignty.

Open access to scientific data is creating new risks to data sovereignty, which is one of the causes of the conflicts between China and the US. Donald Trump and his administration have accused China of failing to share its samples of COVID-19 with other countries (Riley-Smith). The cause of this conflict is data sovereignty, which is essential to technological sovereignty. In the age of AI, data sovereignty is a sine qua non condition of sovereignty.

3. International Instruments for Ethics in Artificial Intelligence

AI technologies are growing faster than international laws. Thus, international laws, such as those concerning the use of force and defense, need to be revised, especially to ban AWS or the future of humanity will be decided by state actors. One of the challenges and a threat to international law, especially to international humanitarian law is an assurance that the commander (human or no human) in the use of weapon systems can and will respect core legal obligations (Davison 3).

The most important challenges imposed by the malicious use of artificial intelligence (MUAI) are not limited to the modernization of some fundamental principles of international law. The vital challenge is "how this law will be interpreted by non-humans" (Abhivardhan 5). Thus, urgent reform of the UN Charter is needed. International law needs to be updated to consider all aspects of AI and its implications, including automation, personhood, weapons systems, control, and standardization. Burri argued that international law must be reviewed as AI entities possess legal personalities (95). The proliferation of AI entities demands that international law reassess this topic, but "neither international law nor will the work of international lawyers be automated because the data remains too limited for AI to learn to give a sound legal assessment" (Abhivardhan 5).

AI poses new challenges related to international law and human rights; these challenges are underscored by the prevalence of the phrase "the Age of AI" used in the development of international law regarding a variety of issues. For that reason, there is a need to rethink international law and to include ethical concerns in AI development, which is the only way to ensure security and to face the tension between states. The ethics of artificial intelligence is the response to a new dilemma that demands international society to provide a legal response to the many ethical challenges artificial intelligence creates. International society needs now, more than at any previous time, to consider the ethics of AI. Several intergovernmental organizations are focusing on this topic.

3.1. A Draft Text for the Recommendation of Ethics in Artificial Intelligence

Since November 2019, UNESCO has started to elaborate on the first global standard-setting instrument for the ethics of artificial intelligence in the form of recommendations (UNESCO b). For that reason, it nominated an Ad Hoc Expert Group (AHEG) composed of 24 individuals from different disciplines, representing all the regions in the world. The process includes inclusive and multidisciplinary consultations with a wide range of stakeholders. According to the Draft Text for the Recommendation on Ethics of Artificial Intelligence elaborated by UNESCO "The values and principles included below should be respected by all actors in the AI system life cycle, in the first place, and be promoted through amendments to existing and elaboration of new legislation, regulations and business guidelines. This must comply with international law as well as with international human rights law, principles and standards, and should be in line with social, political, environmental, educational, scientific, and economic sustainability objectives" (UNESCO b). This draft underlines the importance of the right to life, freedom of expression, privacy among other fundamental human rights and it recommends several policy actions to the Member States. Respect, protection, and promotion of human dignity, human rights, and fundamental freedoms are the first value included in this text with precision and explanation in four paragraphs. According to paragraph 13 "The dignity of every human person constitutes a foundation for the indivisible system of human rights and fundamental freedoms and is essential throughout the life cycle of AI systems" (UNESCO b).

The quality of life of every human being should be enhanced throughout the life cycle of AI systems without any violation or abuse of human rights, or the dignity of humans (UNESCO b). During their interaction with AI systems, "persons should never be objectified, nor should their dignity be undermined our human rights violated or abused" (UNESCO b), this is also extended to "Human rights and fundamental freedoms must be respected, protected, and promoted throughout the life cycle of AI systems." (UNESCO b). Based on the gender approach, the draft includes three other values: environment and ecosystem flourishing, ensuring diversity and inclusiveness, living in harmony and peace. These values are linked to 10 principles which are as follows: proportionality and do no harm, safety and security, fairness and non-discrimination, sustainability, privacy, human oversight and determination, transparency and explainability, responsibility and accountability, awareness, and literacy, multi-stakeholder, and adaptive governance and collaboration. To operationalize these values and principles, the draft described policy actions in 10 areas such as ethical impact assessment should which is important to identify AI's impacts on human rights.

Recommendations concerning gender, data policy, education and research, health and wellbeing, and other policy areas are included in this first international instrument for ethics in AI. Regional organizations are also focusing on this issue. This has led us to question the efficiency of some regional mechanisms, especially when the topic is as universal as AI and its implications on human rights.

3.2. OECD Principles on Artificial Intelligence

On 22 May 2019, the OECD adopted a recommendation on AI (OECD b). The OECD's Recommendation identifies five value-based principles and human rights. Well-being, democratic values, and the respect of the rule of law are at the center of these principles (OECD b). These values are based on the international legal framework and international human rights law. Though well-intentioned, the OECD values proved convoluted, having combined in each a number of individual values. Furthermore, the links between the five values were not made clear and the end result was that their impact was not successful. For example, OECD prioritizes inclusive growth and sustainable development which are ways to ensure wellbeing. The first value is a combination of inclusive growth which is not a goal but a way to ensure wellbeing. It is required to rethink the organization of these values to clarify the difference between the values.

Consistent with these value-based principles, the OECD also provides five recommendations, which underlines the importance to empower people with AI skills and support workers in jobs that will employ AI (OECD a,). While these five recommendations address topics such as fostering trustworthiness in AI, empowering the public and private partnership, they neglect to address the various fields influenced by AI systems. Ambiguity characterizes the link and connexion between the five recommendations and values-based principles. These recommendations should be addressed first to member states and then through them to other stakeholders. While the recommendations are sometimes addressed to public and private sectors and citizens, such entities are not members of OECD and should not be addressed directly. Only member states should be addressed regarding the recommendations (according to the international organizations' constitutions) and then depending on their decision, they can choose to harmonize their national legislation and address the recommendations with other actors (public and private sectors, civil society, academia, individuals, etc.). Such a process emphasizes the legal importance of the text where recommendations serve as non-binding guidelines. International society needs an international instrument with powerful legal value.

In addition to the recommendations, the OECD also implemented an AI Policy Observatory, which aims to offer guidance to countries who need assistance enacting these principles by encouraging, nurturing, and monitoring the responsible development of trustworthy AI systems for the benefit of society. Only 40 countries have adopted these principles including 36 OECD member countries, and six non-member countries (OECD c).

3.3. Artificial Intelligence ethics guidelines

In 2019, the Council of Europe created an Ad Hoc Committee on AI (CAHAI), which is working on "the feasibility and potential elements based on broad multi-stakeholder consultations, of a legal framework for the development, design, and application of artificial intelligence, based on Council of Europe's standards on human rights, democracy and the rule of law" (Council of Europe 1) . Several international organizations are working on rules and legal frameworks related to the ethics of AI, such as the European Commission's High-Level Expert Group on AI (AI HLEG), which produced the DRAFT Ethical Guidelines for Trustworthy AI.

According to the first draft of the AI ethics guidelines prepared by the AI HLEG, "ethical purpose is used to indicate the development, deployment, and use of AI which ensures compliance with fundamental rights and applicable regulation, as well as respecting core principles and values. This is one of the two core elements to achieve Trustworthy AI" (AI HELG 7). The goal of this initiative is to prepare European countries for the tangible and intangible impact of artificial intelligence, including socioeconomic changes, a goal conditioned by European values and ensured by a European ethical and legal framework. Fundamental legal reforms and new policy actions are needed that include the integration of all the stakeholders. The EU is based on a constitutional commitment to protect the fundamental and indivisible rights of human beings as cited in Articles 2 and 3 of the Treaty on the European Union (European Union 5) and the Charter of Fundamental Rights of the European Union (European Parliament 9).

Ethics in AI is reflected in their statements regarding principals, values, and rights. Their starting point is trustworthiness, being the cornerstone that guarantees human rights in an age of artificial intelligence. While the COVID-19 pandemic has accelerated the use of AI in several fields, such as healthcare, manufacturing, and aviation, exacerbating the replacement of peoples' jobs by machines, the increase in the use of AI has additionally exposed humanity to new threats linked to the safety of AI systems. An increase in the importance and impact of data science for trustworthiness in AI is now considered a key in combating COVID-19.

Other global and regional instruments focus on the application of AI in a human-centered approac., For example, the G20 AI Principles were adopted by the G20 Trade Ministers and Digital Economy in June 2019. The principles are drawn from the OECD recommendations on AI. The goal was to include a human-centric approach to AI, which is the only way to guarantee human rights and democracy in the age of AI. According to these principles, trustworthiness in AI is at the center, and it needs the contribution of all stakeholders. Trustworthiness is the first principle cited because it is considered a cornerstone to ensure human rights, democracy, and sustainable development. As stated in the principles, "AI actors should respect the rule of law, human rights, and democratic values, throughout the AI system lifecycle. These include freedom, dignity and autonomy, privacy and data protection, non-discrimination and equality, diversity, fairness, social justice, and internationally recognized labour rights" (G 20). This document is also a call for action, and it contains recommendations that require the engagement of all stakeholders. Part of the document is dedicated to solutions and policy actions that can be adopted by states, and it underlines the importance of international cooperation in this field.

A more recent example of principles for AI ethics was the G7's (2018) Charlevoix Common Vision for the Future of AI adopted in Charlevoix, Canada in June 2018 by the leaders of Canada, France, Germany, Italy, Japan, the United Kingdom, and the United States. It contains 12 commitments. This vision recognized that "AI that fosters economic growth, societal trust, gender equality, and inclusion depends on a predictable and stable policy environment that promotes innovation". Several actions are recommended to the state members based on "ethical and technologically neutral approaches" as stated in the first commitment of this vision.

One final example of such guidelines includes the declaration by the African Union's Working Group on AI, stated by Sharm El Sheik as being adopted by African ministers responsible for communication and information and communication technologies (CICT) in Egypt on 26 October 2019 (African Union). This important legal framework confirms that international society is dedicated to the importance of ethics in AI, including the development of rules and strategic actions to face challenges imposed by AI and the importance of updating international law in the age of AI.

4. Conclusion

In the age of AI, new reforms are needed at different levels considering the changing identity of international society with the emergence of new actors, particularly transnational corporations who have invested in AI more than some States. In the same context, AI is facilitating the creation of a new international order as indicated by the race towards AI in many fields and the large investment of states in these technologies. An effect of this is that international public law will also be influenced by future changes related to AI and the "new Westphalia system" that will be established in the next few years. COVID-19 is accelerating the transition of international society to a new world order characterized by new powers, new faces of war, new forms of slavery, new colonialism, new tools, and a new generation of human rights. Consequently, rethinking international human rights law and international humanitarian law is an urgent task to reduce the risk of AI and to protect human rights in peace and war.

In the same context, it is necessary to update the UN Charter in light of the appearance of these new concepts and phenomena linked to advanced technologies (including AI), which are being used as unconventional weapons. The creation of a multilateral convention under the UN on ethics of artificial intelligence is also needed. New international and national mechanisms are needed to enhance international bilateral and multilateral cooperation on the ethics of AI. The creation of special grants for small countries, to reduce the technological divide between the Global North and South, is a way to face AI implications on human rights. Together with this, the provision of grants in countries, between rural and urban regions, will also contribute to addressing digital divides within states.